Summary of Work: Our research efforts encompassed two general areas: (1) The modulatory effects of bilayer lipids on the structural reorganizations of integral membrane proteins, and (2) the instrumental development and applications of vibrational Raman and infrared spectroscopic imaging techniques. (1) Our interest in characterizing the sizes and formation of fluctuating lipid microdomains within biomembranes, using vibrational infrared and Raman spectroscopy, were focused on lipid cluster formation, aggregate size and their modulatory influence on induced conformational changes occurring within integral membrane proteins. In particular, the compressibilities of systems composed of various lipid microdomains were correlated with intramolecular protein rearrangements. Monomeric bacteriorhodopsin, reconstituted in both multilamellar vesicle systems and single shell vesicle assemblies, was used as a model system to demonstrate the effects arising from the lateral compressibility properties of these quantified lipid microaggregates. In particular, the motional characteristics of this protein's transmembrane helices were monitored by the infrared spectroscopic half-widths of the characteristic Amide I vibrations. This vibrational mode reflects, through heterogeneous broadening mechanisms, the mobilities of the trans-membrane alpha helices during the integral membrane protein reorganizations. To study spectroscopically specific bilayer lipid chain order/disorder properties within the membrane microdomains, appropriate lipid acyl chain deuteration was required to allow the vibrational dynamics of the chain moieties to be monitored. Binary mixtures of saturated chain phosphatidylcholines were specifically examined. Various spectroscopic splitting patterns of the methylene bending modes allowed a determination of lipid microdomain size in terms of the number of acyl chains constituting a given lipid cluster. The compressibilities of the lipid assemblies were determined both isothermally and adiabatically. An infrared diamond anvil cell was used to measure bilayer isothermal compressibility. Pressures were defined by monitoring the spectra of a pressure transducing material, while volume changes were measured directly. Adiabatic compressibilities of the lipid dispersions were determined by ultrasonic velocimetry in which the thermotropic response to the velocity of sound is measured. In examining binary lipid mixtures, microdomain sizes were found to be functions of the lipid mole fractions constituting the system. Specifically, the lateral compressibilities of the binary systems and integral membrane protein reorganizations were governed by the effective domain sizes defining the assembly. (2) Considerable emphasis was placed on enhancing our mid-infrared spectroscopic chemical imaging microscopy techniques by combining step-scan and continuous scanning interferometry with state-of-the-art infrared sensitive two-dimensional focal plane array and linear array detectors. The integration of high performance digital imaging with noninvasive, high resolution infrared spectroscopy allows a visualization of the spatial distribution of distinct chemical species in a variety of host environments. The power of the technique is also manifest in the simultaneous acquisition of an infrared spectrum for each spatial location. As one example of the utility of the infrared imaging technique in diagnostic pathology, we applied this approach to both supervised and unsupervised (observer independent) prostate histopathology involving large numbers of tissue samples examined in the form of tissue microarrays. This infrared spectroscopic method eliminates the necessity for chemically stained tissue. Further, the high throughput approach inherent in the use of tissue microarrays allows the efficient and effective acquisition of a sample?s vibrational spectral signature for pathologic analyses of biopsied specimens representative of controls, prostatic intraepithelial neoplasia, benign prostatic hyperplasia and adenocarcinoma. We specifically demonstrate the application of automated histologic segementation for a series of archival tissue samples; well-defined tests of statistical significance were incorporated. This approach demonstrates that histopathologic changes can now be defined by biochemistry-based, objective spectroscopic criteria that do not necessarily require a pathologist's intervention or interpretation. In these examples, our imaging instrumentation incorporated highly sensitive linear array and focal plane array detection for rapidly recording hypercube spectral data. For spectroscopically elucidating the various histologic features present in prostate tissue, extraordinarily large spectral training sets and appropriate spectroscopic metrics were developed for distinguishing ten morphological entities occurring in this specific tissue. Both the use of the Maximum Gaussian Likelihood Method and, separately, a probabilistic classification model, allowed an objective, automated delineation of the ten histologic categories to be correct to the order of 95-99%. Fine tuning of the tissue segmentation process was developed. Additionally, receiver operating characteristic curves were used to explore relationships between sensitivity and specificity of the high throughput, spectroscopic delineations for distinguishing adenocarcinoma. These procedures are entirely compatible with current tissue processing procedures. With regard to our general infrared imaging instrumentation, a number of enhancing features were made in the optics, in detector configurations, and in data collection paradigms. In particular, we have implemented and utilized a generalized form of interferometric rapid-, or continuous, scan infrared spectroscopic imaging (to be distinguished from interferometer step-scan approaches) for utilization with any type of focal plane array detector to image nonreversible dynamic events in the order of seconds. Further, we are the first group to implement time-resolved Fourier-transform infrared spectroscopic imaging which permits, for example, the visualization of repetitive dynamic processes with half lives on the order of tenths of milliseconds. The examples used in this case involved specified polymer dispersed liquid crystalline composites. In contrast to step-scan interferometer approaches to examine molecular dynamics, we demonstrated the ability to monitor the dynamics of multilamellar lipid bilayers using continuously scanning Fourier transform infrared spectroscopic imaging techniques. The spatially and temporally resolved multilamellar images allowed direct and simultaneous determinations of various physical and chemical properties of the assemblies, including the main thermal gel to liquid crystalline phase transition, comparisons of vesicle diffusion rates in both phases and the variation in lipid bilayer packing properties between the inner and outer lamellae defining the vesicle. Since Fourier-transform infrared spectroscopic imaging data sets are extraordinarily large, data processing becomes the limiting step in visualizing sample heterogeneity and temporal profile evolution. We adapted the Gram-Schmidt vector orthogonalization procedure to interferogram space to provide a significant time saving advantage in processing of one to two orders of magnitude in comparison to conventional spectral processing.